Document

Andrew Gowasack: Hello everyone. My name is Andrew Gowasack. I am one of the two co-founders of Trust Stamp and the President of Trust Stamp Group. And with me today is Gareth Genner, my co-founder, who serves as CEO. While we both do whatever is needed, generally, I oversee sales, marketing, and operations, while Gareth works on strategic partnerships, intellectual property, and legal and capital market compliance.

I'm based in our Atlanta HQ and Gareth is based in our Malta HQ, which gives our working days extended time zone coverage, which is helpful as our team of 80-plus people is spread across four continents now.

Before we dive into the conversation, if you have any questions on the topic that we discuss, please feel free to post them in the message box. The post will only be seen by you and our moderator. So, if you provide your email address and the question's not relevant to today's topic, or we don't have time to address them today, we'll contact you later this week.

Gareth Genner: Thank you, Andrew, and thank you to everybody who's joined this call. At the end of Q1, we put out a written year-to-date brief, and although many investors appreciated the breadth, the feedback was generally that it was too much information in a single document, and it would be much better if we had a conference call.

So, here we are today. The timing conveniently also allows me to talk about the Sovereign Technology Center. As we will be sharing a lot of forward-looking information today, we will also issue a transcript on Monday. But I am required to start—boringly, I'm afraid—by stating that the statements on this call that are not based on historical fact are forward-looking statements, including within the meaning of the Private Securities Litigation Reform Act of 1995, and the provisions of Section 27A of the Securities Act of 1933, as amended, and Section 21E of the Securities Exchange Act of 1934, as amended.

The information in this presentation may contain forward-looking statements and information related to, amongst other things, the company, its business plan and strategy, and our industry. These statements reflect management's current views with regard to future events-based information and are subject to risks and uncertainties that could cause the company's actual results to differ materially from those looking from those contained in the forward-looking statements.

Investors are cautioned not to place undue reliance on the forward-looking statements, which speak only as of the date on which they are made. And the company does not undertake any obligation to revise or update these forward-looking statements to reflect events or circumstances after this date or reflect the occurrence of unanticipated events.

Andrew Gowasack: Thanks, Gareth. And I appreciate you handling the legalese. I know that you're excited to talk about the Sovereign Technology Center, but could you please start by talking about the US Financial Services customers that we've been, have been the traditional mainstay of our business?

Gareth Genner: Certainly. With a six-year contract extension in 2025, our S&P 500 bank partner continues to be the bedrock of our US Financial Services business, and their commitment to innovation constantly pushes us to bring new value by way of product innovation that benefits all of our clients.

We currently anticipate gross revenue from that relationship of around $2.8 million in 2026 and $3 million in 2027, with continued growth in future years. Their move to operating on our low-code orchestration layer has now contributed to a higher gross margin on the services that we provide.

In addition, we have now enrolled over 100 financial institutions on our orchestration layer via our channel partnership. And our transaction volume and revenue from usage has continued to grow. It is consistently lower than we would hope, as even after paying an onboarding fee, the financial institutions typically initiate usage slowly and then in progressive increments of their customer base.

Also on the orchestration layer, we have one of America's largest life insurance companies, together with a number of healthcare, fintech, and non-financial enterprises, and we are in the process of complementing the existing banking channel partnership by integration with another of the US "Big Three" banking platforms. We anticipate that integration should be complete by the end of this quarter.

We've recently added our first credit union to the orchestration layer. They are using our new standalone AMVA product offering, and we believe that there is a significant opportunity to replicate that offering for other small institutions that cannot afford or do not want a broader range of services. We will start marketing that standalone product before the end of this quarter.

Despite the slow build-up of new customer usage volumes, revenue from the orchestration layer is sticky, and we will continue to work on growing that vertical as a stable provider of long-term annual recurring revenue in parallel to our newer revenue channels.

Andrew Gowasack: Thank you, Gareth. Next question: over the last year or so, you've talked about our work in Africa, and in the business report in April, you said that you anticipated being able to share information of potential revenue in Q3. I know that we're not in Q3 yet, but can you tell us today?

Gareth Genner: Certainly. I know everybody is keen to have an update. We are actively working with Africa's largest Telco, and as we make initial product deliveries to them, the range of services they are asking for continues to grow.

We are currently estimating a minimum of 2 million in revenue from that engagement by the end of 2027, with the potential for substantial increases based on the client's speed of rollout and growing geographic deployment. This is an engagement that could well build to 3 million to 5

million of annual recurring revenue over time. We also have a second, smaller telecom under negotiation and hope to have it contracted in Q4 of this year.

Government relations in Ghana are very slow to move from MOU to contract, and then revenue. We currently have a pipeline of about 1.8 million in revenue from contract negotiations for the balance of 2026 and 2027. However, those projects include long-term revenue shares, which, if successful, can also provide an eight-figure annual recurring revenue.

In Nigeria, entirely separate to our Telco engagement, we have multiple government contracts under negotiation with potential revenue for the balance of 2026 and 2027 of up to $20 million US dollars. Again, many of those contracts are designed to also provide a long-term revenue share.

For our planning purposes, we are, however, aggressively discounting that pipeline to 4.6 million for the balance of 2026 and 2027. Frankly, the long-term revenue projections provided by potential customers, governments, and intermediaries in the African market are at such a high level that they would distort our revenue projections. Therefore, we do not include them in our pipeline, but we will be delighted if they prove to be in a realistic range.

Andrew Gowasack: So, shifting markets slightly, Gareth. With the assistance of the Tokyo Metropolitan Government, we've opened an office in Tokyo with a single team member. How has that progressed?

Gareth Genner: Well, the Metropolitan Government and other Japanese agencies have been incredibly helpful in opening doors with large Japanese corporations to allow us to start the lengthy process of building trust, typically through pilot tests, and we're now seeing the fruits.

A couple of months ago, we announced an agreement with Digital Platformer, a provider of software infrastructure for banks and other financial institutions. And we also republished the announcement of the successful completion of testing by INTEC, a highly respected provider of public key infrastructure in Japan, with the target of live deployment by the end of 2026.

This week, we were delighted to be informed that SEIKO Systems, the information technology and systems integration arm of Japan's publicly traded SEIKO Group Corporation, publicly announced successful testing of our technology and their intent to leverage it in a 2026 rollout for multiple use cases.

These include admission authentication in entertainment venues, fraud prevention in financial and payment services, facilities management and security, academic examinations and online certification, digital wallet management, fan club membership authorization, and identity verification for general online services.

If you've never taken a look at the large Japanese corporations, they have a phenomenal number of different businesses in the group, which is reflected by the number of use cases.

Now, we believe that a flagship client like SEIKO Services will provide credibility to bring on other Japanese enterprises once the first product is live in the market. And we see the Japanese market becoming an increasingly important source of opportunities for us.

Now, I know the shareholders would love to see our issuing focused PR for scale opportunities such as this, but the Japanese customer requirement has been that they issue any PR as and when they decide, and we can then amplify what they have said on social media.

While that's a little frustrating, it's obviously important that we maintain that trust by complying with their preferred communication practices. So, you won't see traditional PR releases from us, but you will see PR pieces from our customers being reflected by us in our social media accounts.

Andrew Gowasack: That's really helpful context, Gareth. Another focus announced in 2026 is the stablecoin-focused "Wallet of Wallets," or WoW, as we have trademarked it and called it in-house. There have been limited announcements about what that project or about that project in recent months. Do you mind sharing where that's at?

Gareth Genner: Well, to remind everybody, WoW is a revolutionary Wallet of Wallets structure that does not store user biometrics, usernames, passwords, or recovery keys, but generates account credentials by applying a unique user algorithm to their tokenized facial biometrics.

The beauty of WoW is that there is no honeypot of data and nothing to lose or forget, nothing to be fished, and it eliminates the nightmares of lost and stolen crypto assets. WoW has been focused on the stablecoin market, and I would mention we also hold the "Stable Key" trademark for that marketplace.

The Genius Act that passed in July 2025 prohibits a stablecoin issuer from paying interest or any yield directly to holders. But crypto exchanges and affiliated platforms may still be able to pay rewards to customers who hold stablecoins on their platforms.

Banks argue this is an obvious loophole: an issuer could share reserve income with an exchange, then advertise a 4–5% reward to the customer. Economically and practically, that could look very similar to an interest-bearing bank account, but without bank capital, community lending obligations, or deposit insurance assessments.

The current uncertainty which has held up our rollout of WoW relates to the proposed Clarity legislation and the balance between legacy financial institutions and the stablecoin issuers who are battling as to whether Congress should prohibit or substantially restrict stablecoin yield and rewards paid indirectly through exchanges and other intermediaries, such as the WoW Wallet.

If we are able to offer a reward structure at the wallet level, that will substantially enhance the traction for issuers to promote our wallet and users to subscribe to it. If we're not able to offer a reward structure, core values based on security, privacy, and ease of use will remain key selling points.

Although we've not been able to roll out the product because of the legislative uncertainty, we have participated in multiple industry events in the US and have been building a prospect list of interested institutions.

Unfortunately, my guess as to when the legislative logjam will clear is no better than anyone else's, although I do fear it might not be until after the midterm elections.

Andrew Gowasack: Thanks for sharing those updates, Gareth. And now we can get to what's had a lot of excitement amongst our team with our partners about the Sovereign Technology Center initiative. It's a major initiative which could quickly outpace all of our other activities. Can you please explain what that's about?

Gareth Genner: Absolutely. I've been looking forward to this one. The initiative is a major opportunity for the company, as Andrew says. It combines responses to distinct but overlapping market pressures to create a comprehensive and unique offering that opens both new markets and revenue streams for us.

That said, it is founded upon a decade of experience building proprietary AI microservices and was strengthened by our strategic acquisition of Lex Verify in February to augment our expertise specifically in working with large language models.

The first issue we identified is that as we have worked in Africa, we've perceived pushback on the concept of our processing or storing data as an enterprise that's ultimately US-controlled. Now, that's regardless of contractual provisions, establishing local subsidiaries, and the in-country location of the data storage and processing. We've also been excluded from a number of opportunities in the EU on the same basis.

A flashpoint in both Africa and the EU has been the US Cloud Act, under which cloud services providers must comply with US legal obligations to preserve, back up, or disclose customer communications and records in their possession, custody, or control, regardless of whether those records are located inside or outside the United States and controlled by the company or a subsidiary.

The third issue is more broadly technology sovereignty. This means that countries should, to the greatest extent realistic, have the means to ensure continuity for technology used in key areas of their economy. As the adoption of and reliance on AI tools grows, the potential catastrophes from their non-availability grow more severe and more realistic. And certainly within the European Union, this is a subject of public conversation every single day.

The recent US government export controls to suspend access to Fable 5 and Mythos 5 by any non-US national were a wake-up call for many countries. Even though that was subsequently modified, this exposed a potential vulnerability for non-US entities building on US-controlled LLM models.

Many users do not realize that specialized AI tools that they are using are ultimately reliant upon a frontier LLM and, therefore, are vulnerable to geopolitical complications regardless of the jurisdiction in which their supplier is based.

As a collateral issue, it's increasingly being recognized that the training of the large frontier LLM models can be troublesome, based upon historical, policy, political, and cultural perspectives, and even generally accepted ground truths as to science.

And the potential for deliberate data poisoning to manipulate outputs is very real, with countless agencies marketing services to tilt LLM outputs in favor of you, or your company's reputation, or of course, against your competitors.

Andrew Gowasack: Now, the name "Sovereign Technology Center" implies a place rather than a service. Can you elaborate on the choice of name and what will be offered?

Gareth Genner: You are correct. The name Sovereign Technology Center does denote a location, and within that location, a set of services.

The first Sovereign Technology Center is being developed in Gozo, Malta, with support from the Maltese and Gozo governments, and with the intent of deploying European Union funding as the center develops.

Each Sovereign Technology Center will be operated by a standalone subsidiary with the economic benefits flowing to Trust Stamp, but management is constitutionally vested in a local board, making operations and data protection decisions sovereign to the jurisdiction in which the center is based.

The initial Gozo Center will be the home base for custom-built, air-gapped servers that we trained using auditable data inputs and initially offer specialized LLM models focused on healthcare, compliance, tourism, and education, all being areas critical to the Malta economy. As with all of our products, all personal data will be tokenized to protect the data subject.

But the center will not only build, train, and deliver the models for government enterprise and educational use, but will also offer free model use for educators, researchers, and startup enterprises as a catalyst for the local economy. Gozo, and the overall country of Malta, will serve as a testing ground for new community-facing, privacy-first AI products that can then be exported to other markets.

The custom-built STC servers are designed for eco-friendly operation, which is recognized as very important in the export markets that we're addressing. For many of the target jurisdictions, the capital costs and power and water needs of operating a local giga-center are just not economically sustainable, and also morally questionable.

One academic model has estimated water consumption of as much as 3 liters for an AI-generated email in certain African locations, although frankly, that is very high compared to estimates for frontier models operating in US giga-centers.

Regardless, critical LLM services can be provided in the target communities without building a giga-center or impacting existing power and water resources.

At the Gozo Center, we will be running multiple LLM models on multiple server types to demonstrate functionality based upon different levels of available local resources. We'll also operate small language models, and based upon our work for the delivery of AI at the edge, we will even deliver ultra-small models that can operate on a single semiconductor, in line with our selection to participate in the European Union's semiconductor-focused IPCEI program.

The servers in the first technology service in Malta will be entirely powered by rooftop solar panels and will have a zero net carbon footprint.

Using the Gozo Center as a powerful demonstration site, we will offer other jurisdictions an affordable, out-of-the-box Sovereign Technology Center that will not burden local power, will

Andrew Gowasack: Thanks for those details, Gareth. Now, the big, obvious question from most of the people on this call: how big is the specific market for sovereign AI deliveries?

Gareth Genner: Obviously, this is a nascent market. This opportunity is growing by the day, but published estimates, while varying tremendously, are all very large.

One published estimate places global sovereign AI infrastructure—so, this specific product—at 19.2 billion in 2026, rising to 177.1 billion by 2035. While another estimates 24.8 billion in 2026 and approximately 301.6 billion by 2040. However you look at it, the total addressable market is huge.

We will, however, be initially focusing on the EU and Africa, with the initial Sovereign Technology Center in Malta not just serving Gozo, Malta, and the EU, but also acting as an export showpiece for African markets.

African market predictions suggest close to 1 to 1.5 billion in sovereign AI expenditure in 2026, increasing to 4 to 8 billion by 2030, and 12 to 25 billion by 2035.

Andrew Gowasack: So, that's an impressive-sounding addressable market. But aren't we competing with OpenAI, Anthropic, Google, Facebook, and the other global players—not just from the US, but also from China?

Gareth Genner: Well, that's the beauty and timeliness of this opportunity. In our target markets, the governments and corporations are pushing back and even legislating against using the frontier models for all the sovereignty reasons that I explained earlier. Therefore, we are not competing with the large frontier companies.

Our main competition will be a limited number of locally grown solutions, and yes, in some markets, Chinese providers. But many jurisdictions are pushing back against the Chinese providers, just as they're pushing back on the US providers.

Additionally, the delivery and usage costs of our Sovereign Technology Center model are a fraction of the end-cap costs of frontier AI providers, which, as you will see in the press, are causing many companies to start limiting their employees' use of these AI tools.

We can provide an LLM account with no token limitations at a cost of around $2 US dollars per professional user. Depending on local costs, that can be fixed for the long term so that it doesn't need to change as power costs change.

This makes sovereign LLM resources affordable and sustainable when compared to those frontier providers. And again, we're not competing with them, because the whole point is, in the markets that we're looking at, they do not want to put their data in the hands of those frontier providers.

Andrew Gowasack: So, continuing to look at markets outside of Europe and Africa, is there a market for sovereign AI in the US as well?

Gartner predicts that more than 50% of multinational organizations will have digital sovereignty strategies for their company by 2029, compared with around 10% currently.

The drivers for corporate sovereign models include AI adoption costs across their company, privacy regulation, and geopolitical tension, plus, of course, protecting their own data from their competitors.

The need for privacy in that market plays into our foundational philosophies as a company and our experience, and unlike the large LLM operators, we will not be seen as a competitive threat if providing a sovereign LLM to a US corporation.

Now, interestingly, although we have not explored it, I can foresee interest from some US states and even municipalities, given the current tension between federal and state governments regarding data access and usage, and also the potential for the federal government to take ownership stakes in the frontier providers.

Drawing from a wide range of published estimates—again, in a nascent market, these are estimates—the market size for sovereign AI in the US for corporations can reasonably be estimated at around 15 billion in 2026, reaching 48 billion in 2030 and 115 billion in 2035.

Now, we will undoubtedly see the large consulting companies seeking to enter that market to compensate for the consulting revenue they're going to lose from the use of LLM-based services for their customers.

So, while we do see that as a market opportunity, we are going to, in the first place, be focusing on those other markets that I indicated.

Gareth Genner: Yes, but the model will be more in line with our typical software products as opposed to a standalone corporation, and as I said, we'll face heavier competition.

We'll start with customer discovery in the banking and fintech markets and aim to roll out a first US product in Q1 of 2027. That'll be focused on financial services compliance, but we will also reach out to states and municipalities to gauge potential interest at a local government level.

Andrew Gowasack: And how soon do you anticipate revenue from the Sovereign Technology Center initiative? And, you know, how do we anticipate it contributing to revenue over 2026 and 2027?

Gareth Genner: Well, the great news is, we've already delivered our first sovereign LLM for defined compliance tasks. It's designed to run in our client's own data center in Africa without internet access, and has been trained on auditable inputs.

It was designed to avoid our multinational client having to expose PII for its customers to any cloud-based services and to provide guaranteed operating costs, being limited to our fixed license fee and their marginal electricity costs. They have no token costs, as they would with a frontier model.

We anticipate significant contributions to future revenue from sovereign AI deliveries in line with that first delivery, and we will, of course, provide updates as we progress.

Andrew Gowasack: Gareth, I know we're always cautious to provide pipeline numbers given the fallout and delays from initial prospect engagement to revenue, but can you provide any guidance through the end of 2027 based on contracts and mature negotiations?

Gareth Genner: I will. And you're right, we have always been very cautious about this because we have never wanted to put anything out publicly that we couldn't later stand behind.

Our pipeline from existing clients and actively engaged prospects for the US and Africa combined is 42.4 million through the end of 2027. But in our internal projections, we discount that to 12.4 million, which we believe to be very conservative.

Now, that doesn't include revenue from the nascent Japanese engagements, the sovereign LLM contracts, the Sovereign Technology Center, the WoW Wallet, or the scope and volume increases for current engagements from our existing clients, or from new prospects or other prospects not yet classified by us as actively engaged.

But we do believe that this provides some baseline information to understand the growth in revenue that we are anticipating over the next 18 months.

For everybody listening, our web page, truststamp.ai, has today been updated to better reflect the breadth of our business offerings, and I'd invite everyone to visit it. Please note that the investor page and deck will not be updated for another 7 to 10 days, as they are managed by an external agency, and they need time to allow for compliance and legal reviews.

Just checking the question section now... It doesn't look like we have any questions. So, I want to thank you all for joining us today, and we look forward to engaging again at the end of Q3.